Thermodynamic solubility modeling of 2, 2ʹ, 4, 4ʹ, 6, 6ʹ - hexanitrostilbene (HNS)

Abstract

Accurate thermodynamic modeling is essential for chemical process research and development. In this work, four thermodynamic models: UNIFAC, COSMO-SAC, Hansen Solubility Parameter (HSP) and Nonrandom Two-liquids Segment Activity Coefficient (NRTL-SAC) are examined for their utility to estimate the solubility of 2, 2ʹ, 4, 4ʹ, 6, 6ʹ - hexanitrostilbene (HNS) [CAS# 20062-22-0] in pure and mixed solvents. UNIFAC yields qualitatively correct HNS solubility predictions for a few solvents while COSMO-SAC fails to provide reliable predictions. In comparison to the two group contribution-based predictive models, the hybrid correlative and predictive model HSP gives qualitatively correct results for pure solvents while the other hybrid model NRTL-SAC provides qualitative to semi-quantitative predictions for HNS solubility in both pure and mixed solvents. NRTL-SAC is then used to predict the solubilities of HNS and 2, 2ʹ, 4, 4ʹ, 6, 6ʹ - hexanitrobibenzyl (HNBB) [CAS# 5180-53-0], an important byproduct of HNS synthesis, in Shipp reaction media for HNS synthesis. Moreover, n-octanol-water partition coefficients of HNS and HNBB are predicted with NRTL-SAC.